CN107219935A - It is a kind of towards continuous writing Chinese character, support interaction Chinese character input system and method - Google Patents

It is a kind of towards continuous writing Chinese character, support interaction Chinese character input system and method Download PDF

Info

Publication number
CN107219935A
CN107219935A CN201710380769.4A CN201710380769A CN107219935A CN 107219935 A CN107219935 A CN 107219935A CN 201710380769 A CN201710380769 A CN 201710380769A CN 107219935 A CN107219935 A CN 107219935A
Authority
CN
China
Prior art keywords
stroke
word
candidate
input
user
Prior art date
Application number
CN201710380769.4A
Other languages
Chinese (zh)
Inventor
苏统华
刘锦如
刘策
张程亮
高若岳
戴洪良
彭海兵
Original Assignee
哈尔滨工业大学
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by 哈尔滨工业大学 filed Critical 哈尔滨工业大学
Priority to CN201710380769.4A priority Critical patent/CN107219935A/en
Publication of CN107219935A publication Critical patent/CN107219935A/en

Links

Classifications

    • GPHYSICS
    • G06COMPUTING; CALCULATING; COUNTING
    • G06FELECTRIC DIGITAL DATA PROCESSING
    • G06F3/00Input arrangements for transferring data to be processed into a form capable of being handled by the computer; Output arrangements for transferring data from processing unit to output unit, e.g. interface arrangements
    • G06F3/01Input arrangements or combined input and output arrangements for interaction between user and computer
    • G06F3/02Input arrangements using manually operated switches, e.g. using keyboards or dials
    • G06F3/023Arrangements for converting discrete items of information into a coded form, e.g. arrangements for interpreting keyboard generated codes as alphanumeric codes, operand codes or instruction codes
    • G06F3/0233Character input methods
    • GPHYSICS
    • G06COMPUTING; CALCULATING; COUNTING
    • G06KRECOGNITION OF DATA; PRESENTATION OF DATA; RECORD CARRIERS; HANDLING RECORD CARRIERS
    • G06K9/00Methods or arrangements for reading or recognising printed or written characters or for recognising patterns, e.g. fingerprints
    • G06K9/00402Recognising digital ink, i.e. recognising temporal sequences of handwritten position coordinates
    • GPHYSICS
    • G06COMPUTING; CALCULATING; COUNTING
    • G06KRECOGNITION OF DATA; PRESENTATION OF DATA; RECORD CARRIERS; HANDLING RECORD CARRIERS
    • G06K2209/00Indexing scheme relating to methods or arrangements for reading or recognising printed or written characters or for recognising patterns, e.g. fingerprints
    • G06K2209/01Character recognition
    • G06K2209/011Character recognition of Kanji, Hiragana or Katakana characters

Abstract

It is a kind of to be related to the input system and method for a kind of Chinese character towards continuous writing Chinese character, the Chinese character input system and method that support is interactive, there is the problem of input function is limited, interactivity is low and inefficient to solve prior art.The system includes:For the input module for the stroke track for receiving user's input;The corresponding stroke track for replacing stroke of wrong stroke that the stroke track of user's input or collection interaction optimizing module are sent is received for gathering input module, and according to the acquisition module put on collection density collection stroke track;Identify stroke and stroke order and candidate word and candidate character string to composition are given a mark, record scoring information highest candidate word, the identification module of candidate character string;Display module for showing the word string that scoring information highest candidate word, consecutive word are constituted;For interaction optimizing module of the monitoring users to alternative false candidates word, the confirmation of mistake stroke and feedback action.The present invention is applied to the error correction and input of Chinese character.

Description

It is a kind of towards continuous writing Chinese character, support interaction Chinese character input system and method

Technical field

The present invention relates to a kind of input system of Chinese character and method.

Background technology

Information inputs user in Chinese character multiple with overlapping or write the two or more syllables of a word together mode continuous writing, is carried out different from user single Chinese character is write, and " even pen ", " pen by mistake ", " order of strokes observed in calligraphy mistake " etc. occur during writing, and these are unfavorable for input method system identification The situation of character.

" even pen ", is often referred to preceding stroke most end coordinate and is connected with the first coordinate of rear unicursal, during defined herein as user writing Situation about being write in the way of will should be written as the independent stroke of two or more than two to be connected one." pen by mistake " Refer to the but writing that should be written as this unicursal for another stroke." order of strokes observed in calligraphy mistake " is when referring to writing Chinese characters not according to unified The order of strokes observed in calligraphy rule of defined Chinese character is write.

Due to the presence of above-mentioned situation, the part that input method may be caused to input in the candidate character strings that identification model is provided Character is not the character that user wants.Therefore, it is intended that input method can optimize track by local optimized algorithm Either the combination of stroke or it is in optimized selection by the interactive interference of user, to obtain desired candidate characters.Especially when When user is with the overlapping or many numbers of write the two or more syllables of a word together mode continuous writing Chinese characters, if utilizing cursor choosing after candidate word is submitted Select and change many places character, process is particularly cumbersome.If simply by user simply interaction intervention or input method system Automatic Optimal is searched for, and input method can just provide desired candidate characters, then is saved the time of user's input by effective, and is carried Rise the input experience of user.

There is obvious limitation when being interacted with user in some current Chinese character hand-written input systems.When user exists During writing Chinese characters, system is given a mark by Chinese Character Recognition model to the Chinese character of input, then present marking highest Chinese character or The higher several Chinese characters of person's marking are selected for user, are expected that by the Chinese character of this user mutual debug.But mesh Preceding all Chinese character hand-written input systems or input method are the input in units of isolating Chinese character, this input mode pin Mistake often occurs in " even pen " input for user, so most of current hand-written input system and method are all only suitable For in the hand-written input system or method of individual Chinese character.Also there are a small number of input systems towards continuous writing Chinese character and side at present Method, but this method for being directed to continuous writing often influences whether other once the input identification for a Chinese character occur is wrong Input the identification of Chinese character.The shortcoming of prior art mainly have it is following some:

First, in input trajectory aspect, it is impossible to intervene for stroke writing, it is impossible to the track of optimization input or pen The combination of picture, to obtain more accurately candidate characters.

Second, in identification aspect, for occurred in the way of overlapping writing during the multiple Chinese characters of continuous writing " even pen ", The situation that " pen by mistake ", " order of strokes observed in calligraphy mistake " etc. are unfavorable for input method system identification character is not provided with the mode for correcting mistake imitated.

3rd, simple this means of intervention of removing erroneous words is more single, and if the number of characters length of writing and Need the fault of modification relatively more, the writing amount of user will be caused to dramatically increase so that total used time of user writing significantly increases Plus.

The content of the invention

Problems with of the invention in order to solve prior art presence:

First, in input trajectory aspect, it is impossible to intervene for stroke writing, it is impossible to the track of optimization input or pen The combination of picture, to obtain more accurately candidate characters.

Second, in identification aspect, for occurred in the way of overlapping writing during the multiple Chinese characters of continuous writing " even pen ", The situation that " pen by mistake ", " order of strokes observed in calligraphy mistake " etc. are unfavorable for input method system identification character is not provided with the mode for correcting mistake imitated.

3rd, simple this means of intervention of removing erroneous words is more single, and if the number of characters length of writing and Need the fault of modification relatively more, the writing amount of user will be caused to dramatically increase so that total used time of user writing significantly increases Plus.

And then propose a kind of towards Chinese character input system and method for continuous writing Chinese character, support interaction.

It is a kind of towards continuous writing Chinese character, support interaction Chinese character input system, including:

Input module, the stroke track for receiving user's input;

Acquisition module, the stroke track of user's input, or collection interaction optimizing module are received for gathering input module The wrong stroke correspondence of transmission replaces the stroke track of stroke;And according to the point on collection density collection stroke track, remember simultaneously Record the coordinate of point;

Identification module, the set of the point collected according to entering stroke track correspondence identifies corresponding stroke and stroke Sequentially, or according to the stroke track correspondence for replacing stroke the set of the point gathered identifies corresponding stroke, and is replaced Change wrong stroke;And the candidate word and candidate character string that can be constituted to stroke and stroke order are given a mark, scoring information is recorded Highest candidate word, candidate character string (i.e. multiple Chinese characters of the continuous writing of candidate), and all candidate words stroke and stroke Sequentially;

Display module, for showing the word string that scoring information highest candidate word, consecutive word are constituted;

Interaction optimizing module, the information for monitoring the confirmation of false candidates word, and by alternative wrong stroke according to stroke Order is shown;While listening for confirmation of the user to wrong stroke in alternative wrong stroke, and to wrong stroke Feedback action is simultaneously handled feedback action;The feedback action of described wrong stroke include the replacement of wrong stroke, deletion, Merge feedback action and add the feedback action of stroke;

Stroke of the described alternative wrong stroke corresponding to false candidates word, or false candidates word and several preceding times Stroke corresponding to word selection, either false candidates word and stroke or false candidates word thereafter corresponding to several candidate words And its stroke corresponding to several preceding candidate words, several rear candidate words.

Preferably, described acquisition module includes:

Track gathers submodule, the stroke track for gathering user's input;

Point collection submodule, carries out sampling site while the coordinate of measuring point according to sample density to stroke track.

Preferably, described interaction optimizing module includes:

Erroneous words determination sub-module, for confirming to false candidates word, and by alternative wrong stroke according to stroke Order is shown;

Submodule is monitored in action, for confirmation of the monitoring users to mistake stroke in alternative wrong stroke, and right The feedback action of mistake stroke, the feedback action of mistake stroke includes the replacement of wrong stroke, deletion, merges feedback action and add Plus the feedback action of stroke;

Optimize implementation sub-module, handled for the feedback action to user:

If the feedback action of user is modification and adds, receive user and replace stroke or add the stroke rail of stroke Mark, and it is sent to acquisition module;Collection result subsequently is sent into identification module after collection is finished to be identified;

If the feedback action of user is deletes and merged, the stroke trace information for deleting stroke or merging stroke is sent out Give identification module.

Preferably, the input module can receive user equipment input stroke track (such as mouse movement input) and/ Or the touch-control input of user (as touch or finger are slided).

It is a kind of towards continuous writing Chinese character, support interaction Chinese character input method, including:

S101:Input module receives the stroke track of user's input;

S102:The stroke track of acquisition module collection user's input, carries out sampling site same according to sample density to stroke track When measuring point coordinate;

S103:The set for the point that identification module is gathered according to corresponding to entering stroke track, identifies corresponding stroke And stroke order;And the candidate word and candidate character string that can be constituted to stroke and stroke order are given a mark, scoring information is recorded Highest candidate word, candidate character string (i.e. multiple Chinese characters of the continuous writing of candidate), and all candidate words stroke and stroke Sequentially;

S104:Display module shows the word string that scoring information highest candidate word, consecutive word are constituted;

S105:The word string that the candidate word or consecutive word that user shows according to display module are constituted is interacted;

If user directly confirms, the word string that default candidate word or consecutive word are constituted is correct;

If user confirms to the false candidates word in display candidate word, start interaction optimizing module;For example such as Fruit user thinks that candidate word is not target word, just carries out clicking operation to false candidates word therein;

S106:Erroneous words determination sub-module confirms to false candidates word, and by alternative wrong stroke according to stroke Order is shown;Such as erroneous words determination sub-module have received for clicking operation, then confirm false candidates word pair Original position and end position and corresponding storage information in the storage information answered;

Stroke of the described alternative wrong stroke corresponding to false candidates word, or false candidates word and several preceding times Stroke corresponding to word selection, either false candidates word and stroke or false candidates word thereafter corresponding to several candidate words And its stroke corresponding to several preceding candidate words, several rear candidate words;The alternative wrong stroke of display and false candidates word bit Pass is equipped with, such as false candidates word is the first character of input, then selects to show the stroke or mistake corresponding to false candidates word Candidate word and the thereafter stroke corresponding to several candidate words by mistake;If false candidates word is the last character continuously inputted, Then show the stroke corresponding to the stroke or false candidates word and its several preceding candidate words corresponding to false candidates word;If False candidates word is the middle word that continuously inputs, then shows stroke or false candidates word corresponding to false candidates word and its preceding Stroke corresponding to several candidate words, several rear candidate words.Show false candidates word and its several preceding candidate words and/or The stroke corresponding to several candidate words may be related to the probability that mistake occurs in stroke afterwards, can also be big according to precision and screen Small to be selected, the portable terminal device such as mobile phone then corresponds to the less alternative wrong stroke of display, if being directed to PC then The alternative wrong stroke of selection increase display that can be suitably.

Erroneous words determination sub-module is shown as in the alternative wrong stroke of stroke order arrangement, possible some stroke writing Mistake, some either many stroke or has lacked some stroke, or should be that the stroke of one is shown as many strokes;It is dynamic Make to monitor confirmation of the submodule monitoring users to wrong stroke in alternative wrong stroke, and the feedback of wrong stroke is moved Make, the feedback action of mistake stroke includes the replacement of wrong stroke, deletions, the feedback that merges feedback action and add stroke is dynamic Make;

If the feedback action of user is modification and adds, optimization implementation sub-module receives user and replaces stroke or addition pen The stroke track of picture, and acquisition module is sent to, collection result subsequently is sent into identification module after collection is finished is known Not;If the feedback action of user is deletes and merged, the stroke trace information for deleting stroke or merging stroke is sent to Identification module.

Preferably, identification module is to identify that the stroke and stroke of stroke track are suitable according to the coordinate of point in step S103 Sequence;Then the candidate word and candidate character string that can be constituted to stroke and stroke order according to existing Chinese Character Recognition model are beaten Point, record scoring information highest candidate word, candidate character string, and all candidate words stroke and stroke order;

Preferably, if the feedback action of the user described in step S106 is modification and addition, optimization implementation sub-module connects Receiving user to replace stroke or add the stroke track of stroke, and be sent to the processing procedure after acquisition module includes following step Suddenly:

S1071:The wrong stroke correspondence that acquisition module collection interaction optimizing module is sent replaces stroke or addition stroke Stroke track, according to sample density to stroke track carry out sampling site simultaneously measuring point coordinate;

S1072:Identification module identifies that stroke track is corresponding and replaces stroke or addition stroke;And stroke replacement will be replaced Corresponding wrong stroke, or addition stroke is added to the point of addition that user locks, the candidate word reconstituted and candidate Word string, and the candidate word and candidate character string of composition are given a mark again, record scoring information highest candidate word, candidate word String, and all candidate words stroke and stroke order.

Preferably, if the feedback action of the user described in step S106 deletes stroke or merging to delete and merging The processing procedure that the stroke track of stroke is sent to after identification module comprises the following steps:

S1081:Identification module will delete the information deletion of stroke, or merge into some pictures by stroke is merged, again The candidate word and candidate character string of composition, and the candidate word and candidate character string of composition are given a mark again, record scoring information is most High candidate word, candidate character string, and all candidate words stroke and stroke order.

Preferably, step S102 detailed process is as follows:

The stroke track of track collection submodule collection user's input;

Point collection submodule carries out sampling site while the coordinate of measuring point according to sample density to stroke track.

The invention has the advantages that:

First, in input trajectory aspect, stroke writing being intervened, for one or the feelings of strokes write user more Condition, can carry out stroke deletion;Write less for user one or strokes, stroke insertion can be performed.So as to optimize Entering stroke is to obtain more accurately candidate characters.

Second, in identification aspect, for occurred in the way of overlapping writing during the multiple Chinese characters of continuous writing " even pen ", " pen ", " order of strokes observed in calligraphy mistake " etc. are unfavorable for the situation of input method system identification character by mistake, and there is provided effective error correcting system.For should This is that two strokes are mistakenly identified as one, can be split into the local identification carried out after two again;For that should return Belong to one or several strokes of previous or latter character, redistributing for stroke groupings can be carried out, carried out again afterwards Again local identification.

3rd, in interaction optimizing aspect, based on the combination to stroke it is local redistribute after identification, realize to candidate The optimization of character string;This local search can effectively reduce the writing amount of user, realize more efficient error correction, and final improve is used Family input efficiency.Wrong stroke is it also avoid while the error correction efficiency of some Chinese character is effectively improved causes a Chinese Character Recognition Mistake and have influence on other Chinese Character Recognition mistakes caused by other Chinese-character strokes, can further improve modification efficiency, improve and use Family input efficiency.In the presence of the modification situation of an erroneous words during for continuous writing, compared to the correction of existing Chinese character one by one, sheet Input efficiency can be improved more than 30% by the system and method for invention;The multiple erroneous words occurred in particular for continuous writing Modification situation, input efficiency of the invention is higher.

Brief description of the drawings

Fig. 1 is the structural representation of system described in embodiment one;

Fig. 2 is the monitoring users of embodiment five to the confirmation of wrong stroke in an erroneous words and to wrong pen The schematic diagram of the feedback action of picture;

Fig. 3 is the monitoring users of embodiment five to the confirmation of wrong stroke in two erroneous words and to wrong pen The schematic diagram of the feedback action of picture.

Embodiment

Embodiment one:Illustrate present embodiment with reference to Fig. 1,

It is a kind of towards continuous writing Chinese character, support interaction Chinese character input system, including:

Input module U20, the stroke track for receiving user's input;

Acquisition module U21, the stroke track of user's input is received for gathering input module U20, or gather interactive excellent Change the stroke track for the wrong stroke correspondence replacement stroke that module U24 is sent;And according on collection density collection stroke track Point, while the coordinate of measuring point;

Identification module U22, the set of point collected according to entering stroke track correspondence identify corresponding stroke and Stroke order, or the set of the point gathered according to the stroke track correspondence for replacing stroke identify corresponding stroke, and will It replaces wrong stroke;And the candidate word and candidate character string that can be constituted to stroke and stroke order are given a mark, record marking Information highest candidate word, candidate character string (i.e. multiple Chinese characters of the continuous writing of candidate), and all candidate words stroke and Stroke order;

Display module U23, for showing the word string that scoring information highest candidate word, consecutive word are constituted;

Interaction optimizing module U24, the information for monitoring the confirmation of false candidates word, and by alternative wrong stroke according to stroke Order shown;While listening for confirmation of the user to wrong stroke in alternative wrong stroke, and to wrong stroke Feedback action and feedback action is handled;The feedback action of described wrong stroke includes the replacement of wrong stroke, deleted Remove, merge feedback action and add the feedback action of stroke;

Stroke of the described alternative wrong stroke corresponding to false candidates word, or false candidates word and several preceding times Stroke corresponding to word selection, either false candidates word and stroke or false candidates word thereafter corresponding to several candidate words And its stroke corresponding to several preceding candidate words, several rear candidate words.

Embodiment two:Illustrate present embodiment with reference to Fig. 1,

Acquisition module U21 described in present embodiment includes:

Track gathers submodule U211, the stroke track for gathering user's input;

Point collection submodule U212, carries out sampling site while the coordinate of measuring point according to sample density to stroke track.

Other modules and structure are identical with embodiment one.

Embodiment three:Illustrate present embodiment with reference to Fig. 1,

Interaction optimizing module U24 described in present embodiment includes:

Erroneous words determination sub-module U241, for confirming to false candidates word, and by alternative wrong stroke according to pen The order of picture is shown;

Submodule U242 is monitored in action, for confirmation of the monitoring users to mistake stroke in alternative wrong stroke, with And to the feedback action of wrong stroke, the feedback action of mistake stroke includes the replacement of wrong stroke, deletion, merges feedback action And the feedback action of addition stroke;

Optimize implementation sub-module U243, handled for the feedback action to user:

If the feedback action of user is modification and adds, receive user and replace stroke or add the stroke rail of stroke Mark, and it is sent to acquisition module U21;Collection result subsequently is sent into identification module U22 after collection is finished to be identified;

If the feedback action of user is deletes and merged, the stroke trace information for deleting stroke or merging stroke is sent out Give identification module U22.

Other modules and structure are identical with embodiment one or two.

Embodiment four:

Input module U20 described in present embodiment can receive user equipment input stroke track (such as mouse movement it is defeated Enter) and/or user touch-control input (as touch or finger slide).

Other modules and structure are identical with one of embodiment one to three.

Embodiment five:

In order to accurate and clearly describe method of the present invention, portion of techniques concept is carried out first further Grammar definition, be defined as follows:

Track:T

Stroke:S

Point:P is triple, shape such as (x, y, i).Wherein x represents abscissa;Y represents ordinate;X, y are constant;I is knot Beam identification, identify the point whether be a stroke end, i=-1 is expressed as the end of a stroke, and i ≠ -1 represents not to be one The end of individual stroke.

Data structure:G=({ (x, y, i) }, { T, S, P }, P, T)

Production Q={ T → S ∣ TS, S → P ∣ SP, P → (x, y, i) }

It is a kind of towards continuous writing Chinese character, support interaction Chinese character input method, including:

S101:Input module U20 receives the stroke track of user's input;

In step S101, terminal system receives a string of orderly pens of user writing by man-machine interface or input equipment Draw track, and the calling subsequently by its convenient storage.System can handle while these tracks are received when writing by user They paint onto screen.Drawing and the collection of stroke track can be directly realized by under such as windows platform using MFC, Same function can be realized under Android platform by painting canvas.

S102:The stroke track of acquisition module U21 collection user's inputs, sampling site is carried out according to sample density to stroke track While the coordinate of measuring point;

In step s 102, stroke track is collected and finished, and Background scheduling program technic is according to specified sample frequency or density Gather the point on the point coordinates on stroke, collection stroke track to refer to take from stroke a little according to certain collection density, specifically adopt Collection density is not specified, and sets threshold values;The point includes horizontal, ordinate value and the category information of end of identification three;Screen in principle Coordinate is set up mode and not specified, and does corresponding conversion during processing as required.

S103:The set for the point that identification module U22 is gathered according to corresponding to entering stroke track, identifies corresponding pen Draw and stroke order;And the candidate word and candidate character string that can be constituted to stroke and stroke order are given a mark, record marking letter Cease highest candidate word, candidate character string (i.e. multiple Chinese characters of the continuous writing of candidate), and all candidate words stroke and pen Picture order;

In step s 103, method or interface of the terminal system by man-machine interface and by setting in advance obtain identification As a result and by display module it is shown to user.Under such as windows platform, preceding 20 can be drawn in user interface specified location The candidate character string that individual possible candidate characters are constituted, at the same may be also required to draw it is all be considered as single stroke original pen Draw track or draw the stroke track for specifying number by group according to stroke groupings.Can be by obtaining under Android platform The service commitment of system input method specifies number candidate word in candidate frame.It is considered as single pen for needing all of drafting The primary stroke track of picture needs the stroke track of specified number drawn according to stroke groupings, can be painted by painting canvas System, is conducive to the execution of follow-up monitoring action and the acquisition of action executing position.The each pictures size for example drawn all is 40x40 picture, all pictures are drawn in same a line, and origin (0,0) is the screen upper left corner, then is apparent from for any in region (40,0) operation to second image is belonged to the operation in (80,40), thus can corresponds to and obtain operation object.

S104:Display module U23 shows the word string that scoring information highest candidate word, consecutive word are constituted;

S105:The word string that user is constituted according to the display module U23 candidate words shown or consecutive word is interacted;

If user directly confirms, the word string that default candidate word or consecutive word are constituted is correct;

If user confirms to the false candidates word in display candidate word, start interaction optimizing module U24;For example If the user thinks that candidate word is not target word, clicking operation just is carried out to false candidates word therein;

Described to intervene the operation for referring to user in S105, different platform has different operating mode, not specified.Come for PC ends Say, can be " left button is clicked ", " left double click ", " clicking by right key ", the action of " dragging ", can expand;For smart mobile phone or Person's tablet personal computer or for other have a terminal device of touch-screen, can be " pressing ", " relieving ", " clicking ", " double-click ", " long-press ", " dragging ", " scaling " etc. are operated, and can be expanded.

For example, " left button is clicked " is used for determining track initial position to be optimized, " left double click " is used for determining rail to be optimized Mark end position." clicking by right key " or " long-press " is used for deploying the character for specifying writing in units of the stroke that cutting is opened." drag Drag " selected target performs modification, deletion action to different zones, as shown in Figure 2;Mode of operation can be expanded.Acting Cheng Shi, calls identification module.

S106:Erroneous words determination sub-module U241 confirms to false candidates word, and incites somebody to action alternative wrong stroke according to pen The order of picture is shown;Such as erroneous words determination sub-module U241 have received for clicking operation, then confirm to make mistake Original position and end position and corresponding storage information in the corresponding storage information of candidate word;

Stroke of the described alternative wrong stroke corresponding to false candidates word, or false candidates word and several preceding times Stroke corresponding to word selection, either false candidates word and stroke or false candidates word thereafter corresponding to several candidate words And its stroke corresponding to several preceding candidate words, several rear candidate words;The alternative wrong stroke of display and false candidates word bit Pass is equipped with, such as false candidates word is the first character of input, then selects to show the stroke or mistake corresponding to false candidates word Candidate word and the thereafter stroke corresponding to several candidate words by mistake;If false candidates word is the last character continuously inputted, Then show the stroke corresponding to the stroke or false candidates word and its several preceding candidate words corresponding to false candidates word;If False candidates word is the middle word that continuously inputs, then shows stroke or false candidates word corresponding to false candidates word and its preceding Stroke corresponding to several candidate words, several rear candidate words.Show false candidates word and its several preceding candidate words and/or The stroke corresponding to several candidate words may be related to the probability that mistake occurs in stroke afterwards, can also be big according to precision and screen Small to be selected, the portable terminal device such as mobile phone then corresponds to the less alternative wrong stroke of display, if being directed to PC then The alternative wrong stroke of selection increase display that can be suitably.

Erroneous words determination sub-module U241 is shown as in the alternative wrong stroke of stroke order arrangement, some possible stroke Clerical error, some either many stroke or has lacked some stroke, or should be that the stroke of one is shown as many pens Draw;

For example for the modification of stroke groupings defined in this method, deployed first by " clicking by right key " or " long-press " The word for the writing specified, is obtained for fragment packet information.By the original position stroke for clicking specified stroke groupings to be modified Fragments for packet [k], the end position stroke groupings fragment [k+j] of stroke groupings to be modified is specified by double-clicking, and is closed after submitting And be a videoclip element, all fragments to front and rear each chinese character merge afterwards, and part is carried out within this range Re-search for identification.

Confirmation of the submodule U242 monitoring users to wrong stroke in alternative wrong stroke is monitored in action, and to mistake The feedback action of stroke is missed, the feedback action of mistake stroke includes the replacement of wrong stroke, deletion, merges feedback action and addition The feedback action of stroke;

If the feedback action of user is modification and adds, optimization implementation sub-module U243 receives user and replaces stroke or add Plus the stroke track of stroke, and acquisition module U21 is sent to, collection result is subsequently sent to identification module after collection is finished U22 is identified;If the feedback action of user is deletes and merged, the stroke track for deleting stroke or merging stroke is believed Breath is sent to identification module U22.

The process acted for monitoring users, mode more universal is to monitor mouse for windows platform Click on drag motions.Such as enter after interaction optimizing module, being identified module in all stroke tracks of this writing of user recognizes The track that stroke is constituted to be most likely to be a Chinese character is sequentially arranged, and marks from No. m to No. n (m by continuous underscore <N) it is identified module to think to be most likely to be the stroke of a word, the stroke object that user can pull certain position is repaiied to replacement Change region, afterwards replace writing region stroke is write again, finally double-click " modification " region submission identification module from And obtain new recognition result candidate word.Such as wanting that writing " outstanding " word has been write as " dog " word, then need " perpendicular curved by the 3rd Hook " is revised as " pressing down ".As shown in Figure 2, it is necessary first to which the 3rd " perpendicular crotch " stroke object is dragged to replacement modifier area Decontrol, then write new stroke under the control of old stroke, finally double-click " modification " region and submit identification module to obtain new Recognition result candidate word.This modification process, backstage includes for the operation of data structure:Sampling site and added when writing new stroke In new_points arrays, the index for new stroke writing is added in new_strokes arrays, it is right in modify arrays The stroke element positions that should be changed update the index of the starting position in new_strokes arrays.In addition certain can be pulled The stroke object of position is decontroled to perform deletion action again to region is deleted.

Basic data structure, is designed as follows:

A.points arrays:The sequence of point (x, y, i), addition point (x, y, -1) at the end of unicursal.

B.strokes arrays:Stroke array, index of the record (x, y, -1) in Points arrays, i.e. each stroke The index of finishing touch.

C.modify arrays:Array element is corresponded with strokes arrays, and initial value is -1.If corresponding stroke Element is deleted, then mark is revised as -2.To increase stroke in former strokes sequence, then it should be a nature to correspond to numerical value Number start, for the index of stroke elements starting position in new_strokes arrays of insertion.

D.new_strokes arrays:Terminate since being indexed start to index value for -1, be increased stroke.

E.new_points arrays:Newly-increased points arrays, for storing newly-increased point, are added at the end of unicursal Point (x, y, -1).

F. fragment array:The packet of stroke, identifier identification model classification multiple probability it is higher be probably Chinese Character The packet of the stroke constituted is accorded with, several packet assemblings, which are pieced together, draws a Chinese character.

For the intervention of stroke groupings (fragment), then need first to click the stroke object that determines the packet original position and double The stroke object of end position is hit to determine the starting and ending position being newly grouped.The such as writing Chinese characters word string " people of ABCD mono- EFG " (each capitalization English symbol represents a Chinese character), wherein " people " two words should be contained in the character write originally, As a result wrong identification then needs to modify to original stroke groupings (fragment), and carry out local fragment for " big " word Re-recognize.As shown in figure 3, being clicked first on second stroke object " slash " to determine to be grouped original position, then the Double-clicked on three stroke objects " right-falling stroke " and confirm packet end position, change is submitted afterwards.This change can will represent the stroke of " big " It is grouped (Si-1, Si, Si+1) it is revised as representing two stroke groupings (S of " people "i-1) and (Si,Si+1).This modification process, is knowing , be for by [fragment of the previous character " D " of modification original position " one " word constitutes set]+[" people " in other module New fragment obtained by two character changes constitutes set]+[change the piece of the latter character " E " of " people " word at end position Section constitutes set] search that the fragment of composition is carried out again obtains local new candidate characters.For character D and E both sides Search result before fragment is not made an amendment and intervened.

For Android platform, it can be reached and mouse identical effect by the operation of touch-screen finger.

System and method of the present invention has the following effects that:

First, in input trajectory aspect, stroke writing being intervened, for one or the feelings of strokes write user more Condition, can carry out stroke deletion;Write less for user one or strokes, stroke insertion can be performed.So as to optimize Entering stroke is to obtain more accurately candidate characters.

Second, in identification aspect, for occurred in the way of overlapping writing during the multiple Chinese characters of continuous writing " even pen ", " pen ", " order of strokes observed in calligraphy mistake " etc. are unfavorable for the situation of input method system identification character by mistake, and there is provided effective error correcting system.For should This is that two strokes are mistakenly identified as one, can be split into the local identification carried out after two again;For that should return Belong to one or several strokes of previous or latter character, redistributing for stroke groupings can be carried out, carried out again afterwards Again local identification.

3rd, in algorithm aspect, based on the combination to stroke it is local redistribute after search, realize to candidate characters This local search of optimization of string can effectively reduce the writing amount of user, realize more efficient error correction, and final raising user is defeated Enter efficiency.In the presence of the modification situation of an erroneous words during for continuous writing, system and method can will improve more than 30%; The modification situation of the multiple erroneous words occurred in particular for continuous writing, of the invention is higher.

4th, input module U20 can receive the stroke track of user equipment input and/or the touch-control input of user, carry A kind of simpler simple, more hommization, interactivity more good means of intervention is supplied.

Embodiment six:

Identification module U22 is the stroke that stroke track is identified according to the coordinate of point in step S103 described in present embodiment And stroke order;Then the candidate that can be constituted to stroke and stroke order according to existing or self-built Chinese Character Recognition model Word and candidate character string are given a mark, record scoring information highest candidate word, candidate character string, and all candidate words stroke and Stroke order.

Other modules and structure are identical with embodiment five.

Embodiment seven:

If the feedback action of the user described in present embodiment step S106 is modification and adds, optimize implementation sub-module U242 receives user and replaces stroke or add the stroke track of stroke, and is sent to the processing procedure bag after acquisition module U21 Include following steps:

S1071:The wrong stroke correspondence that acquisition module U21 collection interaction optimizing modules U24 is sent is replaced stroke or added Plus the stroke track of stroke, sampling site is carried out to stroke track according to sample density while the coordinate of measuring point;

S1072:Identification module U22 identifies that stroke track is corresponding and replaces stroke or addition stroke;And stroke will be replaced Replace corresponding wrong stroke, or will add the point of addition that stroke is added to user's locking, the candidate word reconstituted and Candidate character string, and the candidate word and candidate character string of composition are given a mark again, record scoring information highest candidate word, candidate Word string, and all candidate words stroke and stroke order.

Other modules and structure are identical with embodiment five or six.

Embodiment eight:

If the feedback action of the user described in present embodiment step S106 deletes stroke or conjunction to delete and merging And the processing procedure that the stroke track of stroke is sent to after identification module U22 comprises the following steps:

S1081:Identification module U22 will delete the information deletion of stroke, or merge into some pictures by stroke is merged, weight The candidate word and candidate character string newly constituted, and the candidate word and candidate character string of composition are given a mark again, record scoring information Highest candidate word, candidate character string, and all candidate words stroke and stroke order.

Other modules and structure are identical with one of embodiment five to seven.

Embodiment nine:

Present embodiment step S102 detailed process is as follows:

The stroke track of track collection submodule U211 collection user's inputs;

Point collection submodule U212 carries out sampling site while the coordinate of measuring point according to sample density to stroke track.

Other modules and structure are identical with one of embodiment five to eight.

Claims (9)

1. it is a kind of towards continuous writing Chinese character, support interaction Chinese character input system, it is characterised in that including:
Input module (U20), the stroke track for receiving user's input;
Acquisition module (U21), the stroke track of user's input is received for gathering input module (U20), or gather interactive excellent Change the stroke track for the wrong stroke correspondence replacement stroke that module (U24) is sent;And according on collection density collection stroke track Point, while the coordinate of measuring point;
Identification module (U22), the set of the point collected according to entering stroke track correspondence identifies corresponding stroke and pen Picture order, or the set of point gathered according to the stroke track correspondence for replacing stroke identify corresponding stroke, and by its Replace wrong stroke;And the candidate word and candidate character string that can be constituted to stroke and stroke order are given a mark, record marking letter Cease highest candidate word, candidate character string, and all candidate words stroke and stroke order;
Display module (U23), for showing the word string that scoring information highest candidate word, consecutive word are constituted;
Interaction optimizing module (U24), the information for monitoring the confirmation of false candidates word, and by alternative wrong stroke according to stroke Order is shown;While listening for confirmation of the user to wrong stroke in alternative wrong stroke, and to wrong stroke Feedback action is simultaneously handled feedback action;The feedback action of described wrong stroke include the replacement of wrong stroke, deletion, Merge feedback action and add the feedback action of stroke;
Stroke of the described alternative wrong stroke corresponding to false candidates word, or false candidates word and several preceding candidate words Corresponding stroke, either false candidates word and stroke or false candidates word thereafter corresponding to several candidate words and its Stroke corresponding to several preceding candidate words, several rear candidate words.
2. it is according to claim 1 a kind of towards continuous writing Chinese character, support interaction Chinese character input system, its feature It is, described acquisition module (U21) includes:
Track collection submodule (U211), the stroke track for gathering user's input;
Point collection submodule (U212), carries out sampling site while the coordinate of measuring point according to sample density to stroke track.
3. it is according to claim 1 or 2 it is a kind of towards continuous writing Chinese character, support interaction Chinese character input system, its It is characterised by, described interaction optimizing module (U24) includes:
Erroneous words determination sub-module (U241), for confirming to false candidates word, and by alternative wrong stroke according to stroke Order shown;
Submodule (U242) is monitored in action, for confirmation of the monitoring users to wrong stroke in alternative wrong stroke, and To the feedback action of wrong stroke, the feedback action of mistake stroke include the replacement of wrong stroke, deletions, merging feedback action and Add the feedback action of stroke;
Optimize implementation sub-module (U243), handled for the feedback action to user:
If the feedback action of user is modification and adds, receive user and replace stroke or add the stroke track of stroke, and It is sent to acquisition module (U21);Collection result subsequently is sent into identification module (U22) after collection is finished to be identified;
If the feedback action of user is deletes and merged, the stroke trace information for deleting stroke or merging stroke is sent to Identification module (U22).
4. it is according to claim 3 a kind of towards continuous writing Chinese character, support interaction Chinese character input system, its feature It is, the input module (U20) can receive the stroke track of user equipment input and/or the touch-control input of user.
5. it is a kind of towards continuous writing Chinese character, support interaction Chinese character input method, it is characterised in that including:
S101:Input module (U20) receives the stroke track of user's input;
S102:The stroke track of acquisition module (U21) collection user's input, carries out sampling site same according to sample density to stroke track When measuring point coordinate;
S103:The set for the point that identification module (U22) is gathered according to corresponding to entering stroke track, identifies corresponding stroke And stroke order;And the candidate word and candidate character string that can be constituted to stroke and stroke order are given a mark, scoring information is recorded Highest candidate word, candidate character string, and all candidate words stroke and stroke order;
S104:Display module (U23) shows the word string that scoring information highest candidate word, consecutive word are constituted;
S105:The word string that the candidate word or consecutive word that user shows according to display module (U23) are constituted is interacted;
If user directly confirms, the word string that default candidate word or consecutive word are constituted is correct;
If user confirms to the false candidates word in display candidate word, start interaction optimizing module (U24);
S106:Erroneous words determination sub-module (U241) confirms to false candidates word, and incites somebody to action alternative wrong stroke according to stroke Order shown;
Stroke of the described alternative wrong stroke corresponding to false candidates word, or false candidates word and several preceding candidate words Corresponding stroke, either false candidates word and stroke or false candidates word thereafter corresponding to several candidate words and its Stroke corresponding to several preceding candidate words, several rear candidate words;
Confirmation of submodule (U242) monitoring users to wrong stroke in alternative wrong stroke is monitored in action, and to mistake The feedback action of stroke, the feedback action of mistake stroke includes the replacement of wrong stroke, deletion, merges feedback action and addition pen The feedback action of picture;
If the feedback action of user is modification and adds, optimization implementation sub-module (U243) receives user and replaces stroke or addition The stroke track of stroke, and acquisition module (U21) is sent to, collection result is subsequently sent to identification module after collection is finished (U22) it is identified;If the feedback action of user is deletes and merged, stroke will be deleted or merge the stroke track of stroke Information is sent to identification module (U22).
6. it is according to claim 5 a kind of towards continuous writing Chinese character, support interaction Chinese character input method, its feature It is, identification module (U22) is the stroke and stroke order that stroke track is identified according to the coordinate of point in step S103;Then The candidate word and candidate character string that can be constituted to stroke and stroke order are given a mark, record scoring information highest candidate word, Candidate character string, and all candidate words stroke and stroke order.
7. it is according to claim 6 a kind of towards continuous writing Chinese character, support interaction Chinese character input method, its feature It is, if the feedback action of the user described in step S106 is modification and adds, optimization implementation sub-module (U242), which is received, to be used Family replaces stroke or adds the stroke track of stroke, and is sent to the processing procedure after acquisition module (U21) and includes following step Suddenly:
S1071:The wrong stroke correspondence that acquisition module (U21) collection interaction optimizing module (U24) is sent is replaced stroke or added Plus the stroke track of stroke, sampling site is carried out to stroke track according to sample density while the coordinate of measuring point;
S1072:Identification module (U22) identifies that stroke track is corresponding and replaces stroke or addition stroke;And replaced stroke is replaced Corresponding wrong stroke is changed, or stroke will be added and is added to the point of addition that user locks, the candidate word reconstituted and time Word selection string, and the candidate word and candidate character string of composition are given a mark again, record scoring information highest candidate word, candidate word String, and all candidate words stroke and stroke order.
8. according to claim 6 or 7 it is a kind of towards continuous writing Chinese character, support interaction Chinese character input method, its It is characterised by, if the feedback action of the user described in step S106 deletes stroke or merge stroke to delete and merging The processing procedure that stroke track is sent to after identification module (U22) comprises the following steps:
S1081:Identification module (U22) will delete the information deletion of stroke, or merge into some pictures by stroke is merged, again The candidate word and candidate character string of composition, and the candidate word and candidate character string of composition are given a mark again, record scoring information is most High candidate word, candidate character string, and all candidate words stroke and stroke order.
9. it is according to claim 8 a kind of towards continuous writing Chinese character, support interaction Chinese character input method, its feature It is, step S102 detailed process is as follows:
The stroke track of track collection submodule (U211) collection user's input;
Point collection submodule (U212) carries out sampling site while the coordinate of measuring point according to sample density to stroke track.
CN201710380769.4A 2017-05-25 2017-05-25 It is a kind of towards continuous writing Chinese character, support interaction Chinese character input system and method CN107219935A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN201710380769.4A CN107219935A (en) 2017-05-25 2017-05-25 It is a kind of towards continuous writing Chinese character, support interaction Chinese character input system and method

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201710380769.4A CN107219935A (en) 2017-05-25 2017-05-25 It is a kind of towards continuous writing Chinese character, support interaction Chinese character input system and method

Publications (1)

Publication Number Publication Date
CN107219935A true CN107219935A (en) 2017-09-29

Family

ID=59945162

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201710380769.4A CN107219935A (en) 2017-05-25 2017-05-25 It is a kind of towards continuous writing Chinese character, support interaction Chinese character input system and method

Country Status (1)

Country Link
CN (1) CN107219935A (en)

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN108630030A (en) * 2018-06-27 2018-10-09 重庆工业职业技术学院 The demonstration equipment of Accounting Course and the demenstration method of Accounting Course

Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN1132089C (en) * 1996-02-20 2003-12-24 夏普公司 Hand-written character input display device
CN102063620A (en) * 2010-12-31 2011-05-18 北京捷通华声语音技术有限公司 Handwriting identification method, system and terminal
CN102156577A (en) * 2011-03-28 2011-08-17 安徽科大讯飞信息科技股份有限公司 Method and system for realizing continuous handwriting recognition input
CN102193707A (en) * 2010-03-03 2011-09-21 上海三旗通信科技有限公司 Improved handwriting multiword input method for handheld equipment
CN104063176A (en) * 2014-06-25 2014-09-24 哈尔滨工业大学深圳研究生院 Handwriting sequence editable continuous handwriting input method and system
US20150169950A1 (en) * 2013-12-16 2015-06-18 Google Inc. Partial Overlap and Delayed Stroke Input Recognition

Patent Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN1132089C (en) * 1996-02-20 2003-12-24 夏普公司 Hand-written character input display device
CN102193707A (en) * 2010-03-03 2011-09-21 上海三旗通信科技有限公司 Improved handwriting multiword input method for handheld equipment
CN102063620A (en) * 2010-12-31 2011-05-18 北京捷通华声语音技术有限公司 Handwriting identification method, system and terminal
CN102156577A (en) * 2011-03-28 2011-08-17 安徽科大讯飞信息科技股份有限公司 Method and system for realizing continuous handwriting recognition input
US20150169950A1 (en) * 2013-12-16 2015-06-18 Google Inc. Partial Overlap and Delayed Stroke Input Recognition
CN104063176A (en) * 2014-06-25 2014-09-24 哈尔滨工业大学深圳研究生院 Handwriting sequence editable continuous handwriting input method and system

Non-Patent Citations (1)

* Cited by examiner, † Cited by third party
Title
郑军: "一种面向字形分析的汉字输入输出处理系统的设计与实现", 《中国优秀硕士学位论文全文数据库 信息科技辑(月刊)》 *

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN108630030A (en) * 2018-06-27 2018-10-09 重庆工业职业技术学院 The demonstration equipment of Accounting Course and the demenstration method of Accounting Course

Similar Documents

Publication Publication Date Title
US10156981B2 (en) User-centric soft keyboard predictive technologies
EP3084580B1 (en) User interface for overlapping handwritten text input
CN103814351B (en) Collaborative gesture-based input language
US9811193B2 (en) Text entry for electronic devices
US7394934B2 (en) Recognition of electronic ink with late strokes
JP4820382B2 (en) How to provide structure recognition in a node link diagram
TWI476613B (en) User apparatus, system and method for dynamically reclassifying and retrieving target information object
US10191889B2 (en) Systems, apparatuses and methods for generating a user interface by performing computer vision and optical character recognition on a graphical representation
RU2702270C2 (en) Detection of handwritten fragment selection
CN104318138A (en) Method and device for verifying identity of user
CN101021850B (en) Word search apparatus, word search method
CN103324425B (en) The method and apparatus that a kind of order based on gesture performs
CN102156608B (en) Handwriting input method for writing characters continuously
CN104090652A (en) Voice input method and device
US7283126B2 (en) System and method for providing gesture suggestions to enhance interpretation of user input
US5479536A (en) Stroke syntax input device
CN105574090B (en) A kind of filtering sensitive words method and system
US20040001649A1 (en) Method and system for displaying and linking ink objects with recognized text and objects
US8634645B2 (en) Method and tool for recognizing a hand-drawn table
CN103226388B (en) A kind of handwriting sckeme based on Kinect
US20030179214A1 (en) System and method for editing electronic images
US8713464B2 (en) System and method for text input with a multi-touch screen
CN101359275B (en) Handwriting input method for digital equipment, handwriting input device and mobile terminal
KR20080094785A (en) Document overview scrollbar
CA2390503C (en) System and method for providing gesture suggestions to enhance interpretation of user input

Legal Events

Date Code Title Description
PB01 Publication
PB01 Publication
SE01 Entry into force of request for substantive examination
SE01 Entry into force of request for substantive examination